Analytical - web analytics for blogmasters

When I was a freshman in college, there were two “bird” courses that were considered not only easy but actually fun to take. One was “The Hi-Fidelity of Physics”, where I saw one of the very first CD players ever made (1981?). The other course was “The Uses and Abuses of Statistics”. This course was about how marketing and PR people abuse statistical data to sell an idea, service or concept. Thinking about the course reminds of something that great American statesman Mark Twain once said: There’s three kinds of lies: lies, damned lies and statistics.

While it’s probably true that the general populace either completely believes (useless) statistics and or shuns any stats altogether, part of the problem is that people often don’t completely understand statistics. Math, for some reason, is an oft-hated subject in North America. (Although this might also be true in other parts of the world.)

However, if math writers were to take the time and clearly explain some of the concepts, maybe more people would likely come to understand and appreciate the value of real statistics, and to see through the useless numbers. One concept that needs to be better understood is correlation.

The easiest way to explain correlation is by example. (But be warned that these are blunt examples.) One example that I read years ago, if I remember it correctly, is that people who live near power lines are more likely to get cancer. (By power lines, I mean the large metal towers that are usually near power transformers.)

Now, real estate properties built near power lines usually have lower value than equivalent properties built elsewhere. Someone with a sufficient amount of money would probably buy a property away from power lines, but usually for aesthetic reasons. Thus, it’s more likely that owners/ inhabitants of properties near power lines are low-wage earners.

Low wage earners, at least in the United States - where the study was conducted - are less likely to follow a proper diet and/or get proper medical attention. Since early detection of cancer is often crucial to avoiding, a low wage earner may be more likely to get cancer, if other required conditions exist.

The real conclusion to be drawn is that poor people are more likely to get cancer. However, without all the details, it appears that there is a correlation between power lines and cancer. The point being is that you must learn to distinguish between correlation and factor. While being poor and eating poorly is likely the real factor in getting cancer, there does seem to be a correlation (i.e., some sort of relationship) between power lines and cancer. (If you want to find out more about this study, read Power Lines and Cancer.

Now, if you’re still with me, how does this all relate to analyzing your weblog’s web statistics? Well, just looking at random web metrics without digging deeper into an analysis may lead you to incorrect conclusions. For example, while there may be a correlation between, say, cooking weblogs and low traffic, that may be an incorrect conclusion.

What do I mean? Well, while cooking weblogs may be less popular in general than, say, a gadget weblog, that doesn’t mean that you cannot generate some worthwhile web traffic. The problem might actually lie in how you are promoting your cooking blog, whether you are tagging it appropriately to fit into specify categories, and notifying suitable blog directories such as Technorati.

So don’t blindly glance at your web metrics and draw conclusions without digging a little deeper.

(c) Copyright 2006-present, Raj Kumar Dash, http://www.mathgurusonline.com/

Technorati Tags: mathblog, math blog, correlation, conclusion, statistical measure, statistics, web metrics, web analytics